Apple Intelligence is Here, But Is It Any Good?

Today on the AI Daily Brief, Apple Intelligence is here, but is it any good before that in the headlines? Meta is developing an AI search engine to get further away from Google and Microsoft. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. To join the conversation, follow the Discord link in our show. Notes. Welcome back to the AI Daily Brief Headlines edition. All the daily AI news you need in around five minutes. If there is one thing that Mark Zuckerberg does not like, it is being beholden to someone else's ecosystem. One of the most demonstrable events in Facebook's recent history was Apple's ad changes back a few years ago that almost killed Meta's model. Zuckerberg has frequently cited that in his drive towards approaching AI in an open source sort of way. Well, according to the Information, Meta is now developing their own AI powered search engine, a move that they see as focused on lessening their reliance on companies like Google and Microsoft. The Information writes that the search engine will crawl the web in order to serve conversational answers about current events to people using Meta's Chatbot. And as stated, the goal is explicitly to reduce reliance on companies like Google and Microsoft who currently provide Meta AI with a data feed about new sports and stocks. The Information Source said the company has been working for about eight months to build a database of information for the search bot. The news aligns with reports from last week that Meta had struck a multi year deal with Reuters to allow them to serve news articles and chatbot answers. Aaron Ng writes, Meta is working on a search engine. I can't recall another time the large companies were sniping each other so aggressively AI is disrupting every moat. Speaking of search, Google Search is expanding its AI Overviews feature, AKA its Perplexity style interface, bringing the feature now to more than 100 countries. This feature was originally launched in May as a US exclusive, and by making the feature part of the default search experience globally, Google claims it will now have more than a billion monthly users. Part of the rollout includes localization, adding languages including Hindi, Indonesian, Japanese, Portuguese and Spanish. Now you might remember that the feature had a rocky launch with the notorious suggestion that a user should put glue on their pizza. Since then, Google has also had to grapple with the issue of how to introduce ads into the experience, and this is something we've talked about a lot. A challenger like Perplexity gets to reimagine their business model from the ground up, where Google has a very significant business that is threatened by these new approaches. Looking around X, you can find both bullishness and skepticism around this. Andrei Berkhoff writes, Google's AI overview kills perplexity. Both hallucinate, but Google is the default search engine for 95% of people and its rag is and will always be either faster or the same quality or better quality for the same speed. SEO coach Kai Cromwell writes, Google search is in a weird spot right now. On one hand, I'm happy the food and travel sites that are impossible to use because of ads and pop ups were clapped. On the other hand, the AI overview is terrible and the SERPs are no more helpful than they were prior to HCU's. I would only add that any intermediate lack of capacity with AI tends to be a fairly quick moment in time. Next up, we were just talking about Meta and open source, but definitions around those words have been a little murky. Well now the Open Source Initiative has published their first definition of what it means for an AI model to be open source beyond just publishing model weights. OSI insists that pertinent details of training data, such as how it was sourced and licensed, must be published. Essentially, the standard is that enough information must be known so that a person could substantially recreate the model, not just use it. OSI EVP Stefano Maffuli said an open source AI is an AI model that allows you to fully understand how it's been built. That means you have access to all the components, such as the complete code used for training and data filtering. He noted that the public should have full usage rights to use and modify the model for any purpose, adding most importantly, you should be able to build on top. Maffuli commented that nailing down a consensus definition of open source AI is particularly important given the state of regulation. The eu, for example, is considering giving special carve outs to open source AI and other global regulators could follow suit. He added that our hope is that when someone tries to abuse the term, the AI community will say we don't recognize this is open source and it gets corrected. Now this is obviously squarely aimed at meta who publish model weights but do not disclose the source of training data. After discussions with the osi, Google and Microsoft agreed to drop their use of the term open source, but Meta is yet to do so. A Meta spokesperson said, we agree with our partner, the OSI on many things, but we, like others across the industry, disagree with their new definition. There is no single open source AI definition, and defining it is a challenge because previous open source definitions do not encompass the complexities of today's rapidly advancing AI models We make llama free and openly available, and our license and acceptable use policy help keep people safe by having some restrictions in place. We will continue working with the OSI and other industry groups to make AI more accessible and free responsibly, regardless of technical definitions. Christian Stout, the Director of Innovation Policy at the Law Econ center, had a really interesting take on this, saying the new OSI definition of open source AI takes far too narrow a view. Treating AI openness as binary, either totally open or not open, rather than as a diverse set of options, fundamentally misunderstands modern AI development and deployment. The reality is that training modern AI systems is incredibly expensive and complex. A definition demanding total openness ignores this reality and will create marginal incentives to keep models fully closed. Even with 100% open code and documentation, most orgs can't replicate training runs that cost millions in computer compute. Having the theoretical right to modify something you can't afford to run isn't going to create practical openness. Thus, when firms release models with limited restrictions, but make the weights freely available, they're arguably creating more practical openness than a theoretically quote unquote pure open source model that no one but other large companies can actually run. The requirement for complete training data documentation is actively harmful to innovation. Further, in fast moving AI research, waiting to document everything perfectly before release means valuable models stay private longer. This isn't just academic government contracts and grants can have requirements related to open source AI. If policymakers adopt OSI's narrow definition, we'll see fewer solutions qualify for public funding and contracts, effectively pushing government AI work towards a tiny subset of mega funded projects. All of these release approaches should be considered open source, just with different terms for different purposes, similar to how Creative Commons offers different licenses for different needs. The OSI definition will reduce openness in AI, not increase it. We need a more practical definition that acknowledges the reality of how AI systems are built and deployed. Now, I tend to be in agreement that we might be in a better position if we had a more broad spectrum of definitions, and that there are a lot of unintended negative consequences for this as well. But what do you think? Let me know in the comments either here on YouTube or Spotify. But for now that is going to do it for today's AI Daily Brief Headlines edition. Next up, the main episode. Today's episode is brought to you by Plum. Want to use AI to automate your work, but don't know where to start? Plum lets you create AI workflows by simply describing what you want. No Coding or API keys required. Imagine typing out AI, analyze my Zoom meetings and send me your insights in notion and watching it come to life before your eyes. Whether you're an operations leader, marketer or even a non technical founder, Plum gives you the power of AI without the technical hassle. Get instant access to top models like GPT4O, Cloudsonnet, 3.5, Assembly AI and many more. Don't let technology hold you back. Check out useplum. That's Plum with a B for early access to the future of workflow automation Today's episode is brought to you by Venice. Venice is a private, uncensored, generative AI app. It accesses open source models to enable text, image and code generation without the fear of being spied on or having your data exploited. Discuss anything with Venice without concern about it being monitored, sold or given to advertisers and governments. Venice is different because your conversations and creations are kept securely within the browser, never stored or accessible by Venice. Unlike other AI apps, Venice won't tell you what's okay to say or not. Venice won't patronize you. It simply provides direct access to machine intelligence. No topics are off limits, no ideas are taboo. With Venice, you're in control of the AI as you should be. Pro subscriptions are available for $49 a year or $8 per month. AI Daily Brief listeners receive a 20% discount on Venice Pro. Visit Venice AI NLW and enter the discount code NLWDAILYBRIEF. That's NLW Daily Brief, all one word. Today's episode is brought to you by Superintelligent. Every single business workflow and function is being remade and reimagined with artificial intelligence. There is a huge challenge, however, of going from the potential of AI to actually capturing that value. And that gap is what superintelligent is dedicated to filling. Superintelligent accelerates AI adoption and engagement to help teams actually use AI to increase productivity and drive business value. An interactive AI use case registry gives your company full visibility into how people are using artificial intelligence right now. Pair that with capabilities, building content in the form of tutorials, learning paths, and a use case library and superintelligent helps people inside your company show how they're getting value out of AI while providing resources for people to put that inspiration into action. The next three teams that sign up with 100 or more seats are going to get free embedded consulting. That's a process by which our super intelligent team sits with your organization, figures out the specific use cases that matter most to you, and helps actually ensure support for adoption of those use cases to drive real value. Go to BeSuper AI to learn more about this AI enablement network. And now back to the show. Welcome back to the AI Daily Brief. This is a moment that we have long been anticipating. Ever since ChatGPT kicked off the race for generative AI, with all the big tech players playing along, Apple has been the laggard in the pack. Now everyone has given Apple the benefit of the doubt as they tend not to be the first to move, but instead try to be the best when they do. And so it sort of followed on that earlier this year we started to get rumors around what was happening with Apple and then we finally got some announcements. Characteristically, at the Worldwide Developer Conference, they announced that they would not be doing artificial intelligence, but instead Apple Intelligence, their very own version of AI. And it was very clear from the beginning that the approach that Apple was going to be taking was not about being the leader when it came to frontier models. In fact, recently we've had reports from inside that company that says that they believe that they are two years plus behind other companies when it comes to their own models. Instead, what Apple was quite clearly trying to do was be the best at productizing AI into a way that normal people would use. Apple Intelligence was AI, but for normies. However, it hasn't been all smooth sailing since then. Almost immediately upon announcing that Apple Intelligence would be coming out alongside the new iPhone and iOS 18, Apple had to walk it back and say that actually it would be out about a month after the iPhone came out. This is a big deal because iPhones still make up a huge portion of Apple's annual revenue, and Apple Intelligence is really being looked at by both the company and by analysts as a reason for people to upgrade as upgrade cycles have slowed. So that is a little bit of backstory and as of yesterday we finally started getting these features. Tim Cook tweets. Apple Intelligence is here. Starting today you can get more done on iPhone, iPad, and Mac with powerful new features including system wide writing tools, a more conversational Siri Intelligence in photos, and so much more. This is the beginning of an exciting new era, the Verge writes. IOS 18's cornerstone software feature is finally shipping. At least the first part of it is the blog points out that Apple Intelligence has been available in developer and public beta builds for the past few months, but this is the first time that it's available in a public release. However, as they point out, the features are still being marked as beta and what is actually available is kind of limited The Verge called it an AI starter kit set of features, writing tools to help summarize notes, change the tone of messages, et cetera, AI summaries and notifications and emails, a new updated Siri with this glowing border around the screen, et cetera. Now what's more is even for this limited set of features, you have to get on a wait list. This has folks none too happy, Ankit Harathy says. I'm on a 16 pro have updated to 18.1 and there's still a waitlist to get access to Apple Intelligence. Jobs would have never let this ship just release 18.1 when it's actually available to everyone. Gabrielle Taibbi writes, you're telling me I downloaded iOS 18.1 to get Apple Intelligence and it's still not even technically out? I had to join a wait list insane Broadly, there's a lot of the sentiment summed up by Mikey here, who says Apple Intelligence is the sloppiest rollout Apple has ever done. Basically, we have issues both of what is available, to whom it's available and how much isn't currently available. Now Apple is saying that more tools are going to start arriving in December. If things go according to plan. That's when you'll be able to use ChatGPT inside of Siri, and Apple's AI camera feature. Visual Intelligence will be able to tell you about the things that are around you. A couple of months after that, Apple says it'll launch priority notifications as well as a major upgrade for Siri, including awareness of what's on your screen and the ability to take action within apps. Now obviously that could be a pretty powerful update given this sort of agentic world we're moving into, but that's still pretty far down the line. The Wall Street Journal reinforced how much Apple Intelligence and the iPhone need each other in a piece titled why Apple's AI Success Hinges on the iPhone, they write. Big tech rivals can distribute AI widely through their services, while Apple's hardware centric business needs its devices as the touch point. And once again, they point out that the phased launch quote unquote complicates the picture. Krish Sankar of TD Cowan wrote in a note to clients, gradual AI feature rollout could be hindering a more pronounced upgrade cycle. The Wall Street Journal points out that the quote concerns have weighed on Apple Stock shares are up less than 5% since the iPhone 16's launch event in early September, lagging behind other mega cap techs as well as the S&P 500. Now, the logic for how this plays out in Apple's favor is clear. The Journal again says as capabilities move from cloud services to edge devices such as smartphones and PCs, Apple has a natural advantage as the world's predominant maker of such products. But there is some question as to whether the current crop of AI smartphones can really deliver powerful experiences. Jeffrey's analyst Edison Lee doesn't think so. In fact, they wrote, to expect an accelerated smartphone replacement cycle now due to AI is premature in our view. So lots of really interesting dynamics going on. But so far I've told you about the reactions to variant franchise technology users and to variant franchise market participants. But what about those normal folks that it seemed so clear that Apple Intelligence was aimed at? It is certainly not a big enough representative sample to really know for sure, but when you start to browse around Twitter outside of tech and finance circles, there are a lot of folks who are really excited about some very basic features. K Brew on Twitter, who describes themselves as a theme park enthusiast and foodie than rather writes, wow, with Apple Intelligence you can erase unwanted people from your photos, and it shows him posing in front of a restaurant having removed a half dozen people or more from the shot when someone tried to make a joke about it. Kabreu said they still need to work on it a little bit more, but it's still very cool technology and there are lots of people using explicitly this feature and discussing how much they like it. So maybe Apple will really have the last laugh for now. I think people are still rightly skeptical, but I certainly wouldn't count them out. That's going to do it for today's AI Daily Brief. Until next time, peace.